Performance

OxiDex delivers exceptional performance improvements over the Perl-based ExifTool through zero-cost abstractions, parallel processing, and native compiled code.

Benchmark Results

The following benchmarks compare OxiDex against the original Perl ExifTool running on identical hardware.

Summary

Scenario	Perl ExifTool	OxiDex	Speedup
Single JPEG Read	116.5ms ± 15.6ms	31.8ms ± 14.1ms	3.7x faster
Batch Processing (1000 files)	1911.4ms ± 171.9ms	197.6ms ± 3.1ms	9.7x faster
Write Operation	200.7ms ± 50.4ms	23.0ms ± 1.6ms	8.7x faster
Format Detection	67.3ms ± 13.9ms	10.4ms ± 3.6ms	6.5x faster

Benchmark Metadata

• Date: 2025-12-10 04:28:14 UTC
• Commit: 0574128

System Specifications

• OS: Linux (Ubuntu 22.04)
• CPU: x86_64 (4 cores)
• Memory: 8GB RAM
• Perl ExifTool: latest version
• OxiDex: version 1.1.0

Live Benchmark Reports

View detailed interactive benchmark results from our CI/CD pipeline:

• 📊 Main Benchmark Report
• 📁 Single File Extraction
• 📦 Batch Processing (100 JPEGs)
• 🎯 Format Comparison
• 🔍 Format Detection
• 📝 Full Metadata Read

These reports are automatically generated by Criterion.rs on every commit to main and include:

• Performance graphs with statistical analysis
• Historical trend comparisons
• Outlier detection
• Detailed timing breakdowns

Key Performance Improvements

1. Single File Operations (3.7x faster)

Zero-cost abstractions and compiled code eliminate Perl interpreter overhead:

• No runtime interpretation
• Direct memory access
• Optimized binary parsing with nom

2. Batch Processing (9.7x faster)

Parallel processing with Rayon leverages all CPU cores:

• Process 1000 files in 197.6ms vs 1911.4ms
• Linear scaling with CPU core count
• Efficient work stealing scheduler

3. Write Operations (8.7x faster)

Efficient binary manipulation and atomic file operations:

• In-place metadata updates
• Memory-mapped I/O with memmap2
• Atomic file writes prevent corruption

4. Format Detection (6.5x faster)

Native compiled code dramatically outperforms interpreted Perl:

• Efficient magic byte matching
• Compile-time optimizations
• CPU cache-friendly algorithms

Reproducing Benchmarks

Comparative Benchmarks (vs Perl ExifTool)

Run the comparative benchmark suite:

# Install prerequisites
brew install hyperfine exiftool  # macOS
# or
sudo apt install hyperfine libimage-exiftool-perl  # Ubuntu

# Build OxiDex in release mode
cargo build --release

# Run comparison benchmarks
./benches/exiftool_comparison.sh

# View results
cat benches/benchmark_results.md

Library Micro-Benchmarks

Run detailed benchmarks for internal operations:

# Run all benchmarks
cargo bench

# Run specific benchmark
cargo bench format_detection
cargo bench jpeg_segment_parsing
cargo bench tiff_ifd_parsing
cargo bench full_read_metadata

View HTML reports:

# macOS
open target/criterion/report/index.html

# Linux
xdg-open target/criterion/report/index.html

# Windows
start target/criterion/report/index.html

Profiling

For detailed profiling information, see Profiling Guide.

Performance Tips

For Maximum Throughput

1. Use batch processing: Process multiple files in one command for parallelism
2. Extract specific tags only: Use -TagName to avoid parsing entire file
3. Binary output: Use -b for binary tag values (faster than formatting)
4. Disable verbose output: Use -q for quiet mode

Example: Fast Batch Extraction

# Extract just Make, Model, DateTimeOriginal from 10,000 JPEGs
oxidex -Make -Model -DateTimeOriginal -csv -r photos/ > metadata.csv

Ongoing Optimization

We continuously monitor and improve performance:

• Automated benchmarks on every commit
• Profiling with samply
• Flamegraph analysis
• Memory allocation tracking

See our optimization strategy for current focus areas.